Tagged articles
Data Efficiency
18 articles · Page 1 of 1
Researchers from Baidu, Renmin University, Tsinghua, and others introduce LoopVL, a recurrent vision-language model that reuses a 1B-parameter language backbone across 128 layer calls, achieving competitive performance with 260x less training data and exhibiting a 'Visual Aha Moment' where attention sharply refocuses on key image regions in the second loop.
Attention MechanismData EfficiencyLoop Transformer
0 likes · 14 min read

MIT CSAIL and Tsinghua researchers introduce GeoPT, a geometric pre-training method that uses synthetic dynamics on 10,000+ 3D shapes to accelerate physics simulation convergence by 2x and reduce required training data by 20-60% across aerodynamics, hydrodynamics, and collision tasks.
Data EfficiencyGeoPTICML 2026
0 likes · 8 min read

Tsinghua-led research shows On-Policy Distillation (OPD) achieves over 70% of full-data performance with just one training example, revealing that data coverage saturates quickly while algorithmic absorption speed becomes the bottleneck.
Absorption RateData EfficiencyModel Training
0 likes · 16 min read

The article breaks down large‑scale neural‑network training into three error components—data, optimization, and architecture—derives a heterogeneous power inequality, and uses it to analyze optimal learning rates, batch sizes, model size, sparsity, and data scaling, linking the results to known laws such as Kaplan, Chinchilla, Step, Microsoft, and DeepSeek.
Data EfficiencyDeep LearningOptimization
0 likes · 28 min read

The article reviews the history, theory, and empirical findings of scaling laws for neural language models, compares the Kaplan and Chinchilla formulations, discusses data‑limited regimes and fitting subtleties, and highlights why careful interpretation and resource allocation are essential for reliable predictions.
Data EfficiencyDeep LearningKaplan
0 likes · 26 min read

The paper introduces Decoupled Residual Denoising Diffusion (DRDD), which reinterprets Gaussian noise as a domain harmonizer and separates residual removal from denoising, enabling more data‑efficient, multi‑task image‑to‑image translation and achieving state‑of‑the‑art results on benchmarks such as All‑in‑One‑5 with limited paired data.
Computer VisionDRDDData Efficiency
0 likes · 14 min read

UniVidX, a unified multimodal framework for video generation and understanding accepted at SIGGRAPH 2026, reformulates diverse video graphics tasks as conditional generation, achieving or surpassing state‑of‑the‑art performance while demonstrating strong data efficiency and cross‑domain generalization.
Data EfficiencyMultimodal Video GenerationSIGGRAPH 2026
0 likes · 10 min read

The article critiques the goal‑driven focus of Generalist's world models, introduces the action‑centric GigaWorld‑Policy architecture that makes video generation optional, explains its three‑stage training pipeline, and presents experimental results showing ten‑fold training efficiency, 360 ms inference per step, and an 83% success rate on real‑robot tasks.
Action‑Centric ArchitectureData EfficiencyGigaWorld‑Policy
0 likes · 11 min read

An AI startup claims that training with only a thousand hours of data produced emergent intelligence and outperformed industry leaders in benchmark tests, prompting a debate over whether this represents a paradigm shift in efficient learning or an overhyped breakthrough requiring further validation.
AIData Efficiencybenchmark
0 likes · 5 min read

In industrial RLHF, repairing bad cases—structural error samples—provides explicit alignment signals that improve model capability far more efficiently than simply increasing data volume, because it teaches the model how to correct mistakes rather than just exposing it to more examples.
Capability ImprovementData EfficiencyRLHF
0 likes · 9 min read

The paper introduces the LIMI framework, which achieves state‑of‑the‑art agent performance on AgencyBench using only 78 carefully crafted samples—outperforming baseline models trained on thousands of examples—by focusing on high‑quality, strategic data construction and demonstrating superior generalization across code, research, and tool‑use tasks.
AgencyBenchAgent AIData Efficiency
0 likes · 11 min read


Amap TechOct 6, 2025 · Artificial Intelligence
World-Env introduces a virtual training sandbox that eliminates physical interaction, dramatically improves data efficiency with just five expert demos per task, and employs a vision‑language model as a semantic judge to dynamically terminate actions, enabling safe, high‑performing VLA post‑training across diverse robotic benchmarks.
Data EfficiencyWorld Modelvirtual environment
0 likes · 9 min read


JD TechMar 12, 2025 · Artificial Intelligence
The article recounts a JD retail engineer’s rapid growth through tackling low‑resource large‑model training, developing a margin‑based dynamic data selection method (DynaMS) that earned an ICLR paper, and sharing practical insights on aligning business needs with cutting‑edge AI research.
AI researchData EfficiencyICLR
0 likes · 11 min read


AIWalkerJan 18, 2025 · Artificial Intelligence
Shanghai AI Laboratory’s InternLM 3.0 upgrade demonstrates that a refined 4 TB token dataset can boost a large‑language model’s performance beyond that of open‑source peers trained on 18 TB, cutting training cost by over 75% while merging regular dialogue with deep reasoning capabilities.
AI-evaluationData EfficiencyInternLM
0 likes · 9 min read


AIWalkerJan 17, 2025 · Artificial Intelligence
Shanghai AI Laboratory’s InternLM 3.0 upgrade demonstrates that refining data quality—measured as intelligence‑per‑token—can replace massive datasets, achieving higher reasoning and dialogue capabilities with just 4 TB of tokens, cutting training cost by over 75 % while approaching GPT‑4‑level performance.
AI researchData EfficiencyInternLM
0 likes · 9 min read


AIWalkerJan 16, 2025 · Artificial Intelligence
InternLM 3.0 (InternLM‑3) upgrades the Shusheng‑PuYu model by refining data to boost "thinking density", using only 4 TB of tokens to surpass peer open‑source models, cutting training cost by over 75% while merging ordinary dialogue with deep reasoning capabilities.
Data EfficiencyInternLMModel Evaluation
0 likes · 9 min read

This report, jointly authored by Qianzhan Industry Research Institute, Shougang Fund CANPLUS, and Huawei Cloud, examines AI large‑model applications across four dimensions—overview, current status with case studies, pain points and solutions, and future trends—highlighting how these models boost production efficiency, balance cost, privacy, and performance, and elevate data’s role in industry.
AIData Efficiencyartificial intelligence
0 likes · 4 min read

This article details Alibaba's large‑scale data model governance initiative, analyzing current data issues, presenting a comprehensive solution—including model digitization, public model sinking, productization, daily governance, and search‑enhancement—and outlining achieved results and future plans to further improve data quality, reuse, and operational efficiency.
Data EfficiencyData ModelingDataWorks
0 likes · 12 min read
